A Quickstart Guide to PyMC3

on YouTube ↗57m

" summary

Step‑by‑step code for building and sampling linear, hierarchical linear, logistic and hierarchical logistic models using PyMC3 with the Theano backend. Demonstrates simulated data and Boston housing, shows concrete issues (pandas frames -> numpy arrays, booleans -> ints, Python2/Unicode string quirks), compares ADVI (very fast) and NUTS sampler (much slower), reports typical R2/accuracy around 75–80%, and explains saving traces for an API. Github repo with the notebooks is provided in the description.

AI-written summary · about

" chapters (14)

1. 0:00PyMC3 introduction
2. 3:35Setting up PyMC3 and basic imports
3. 5:52Linear regression model
4. 8:40Sampling with NUTS
5. 10:23Analyzing results
6. 13:16Real-world data use
7. 15:21Common errors
8. 20:29Posterior distributions
9. 27:24Hierarchical linear regression
10. 31:52Logistic regression & VI
11. 37:40Save and reuse traces
12. 42:25Hierarchical logistic regression
13. 50:43Production tips
14. 54:45Q&A

" related talks

• 44mAn introduction to PyMC3Adrian Seyboldt · PyCon.DE 2017
• 01h 31mProbabilistic Programming and Bayesian Inference in PythonLara Kattan · PyOhio 2019
• 39mProbablistic Programming Data Science with PyMC3Thomas Wiecki · PyData London 2016
• 01h 30mAn introduction to Markov Chain Monte Carlo using PyMC3Chris Fonnesbeck · PyData London 2019
• 42mBayesian Non-parametric Models for Data Science using PyMC3Christopher Fonnesbeck · PyCon 2018
• 01h 27mProbabilistic Python: An Introduction to Bayesian Modeling with PyMCChris Fonnesbeck · PyData London 2022
• 41mAll that likelihood with PyMC3Junpeng Lao · PyData Berlin 2018
• 01h 27mBayesian Time SeriesChris Fonnesbeck · PyData London 2025